def apis(): dt = DumpTruck('/tmp/open-data.sqlite', auto_commit = False) dt.create_table({'catalog':'abc.def'}, 'socrata_apis') dt.create_index(['catalog'], 'socrata_apis', unique = True, if_not_exists = True) socrata_catalogs = filter(lambda x: x[0] == 'socrata', catalogs()) for _, catalog in socrata_catalogs: dt.upsert({ 'catalog': catalog.split('://')[-1], 'apis': count_apis(catalog), }, 'socrata_apis')
def main(): edges = build_network()['edges'] dt = DumpTruck(dbname = '/tmp/open-data.sqlite', adapt_and_convert = True) datasets_in = dt.execute('SELECT * FROM socrata') dt.create_table({'id': 'blah-blah'}, 'socrata_deduplicated') dt.create_index(['id'], 'socrata_deduplicated', if_not_exists = True, unique = True) for dataset in dedupe(datasets_in, edges): dt.upsert(dataset, 'socrata_deduplicated')
def extract_dataset_table_info(): dt = DumpTruck(dbname = '/tmp/table_info.db') dt.create_table({'portal': 'abc', 'id': 'abcd-efgh'}, 'table_info') dt.create_index(['portal', 'id'], 'table_info', unique = True) dt.create_index(['tableId'], 'table_info', unique = False) done = set([tuple(row.keys()) for row in dt.execute('SELECT portal, id FROM table_info')]) for portal in os.listdir('data'): for viewid in os.listdir(os.path.join('data', portal, 'views')): if (portal, viewid) in done: continue d = _dataset_table_info(portal, viewid) if d == None: continue dt.upsert(d, 'table_info')
def get_links(softwares = ['ckan','socrata']): dt = DumpTruck('/tmp/open-data.sqlite') dummyrow = dict(zip(['software','catalog','identifier', 'status_code', 'headers', 'error'], (['blah'] * 3) + ([234] * 1) + ([{'a':'b'}] * 2))) dt.create_table(dummyrow, 'links', if_not_exists = True) dt.create_index(['software','catalog','identifier'], 'links', if_not_exists = True, unique = True) for software in softwares: for catalog in read.catalogs(software): if SOCRATA_FIX.get(catalog, 'this is a string, not None') == None: continue try: for row in _check_catalog(software, catalog): dt.upsert(row, 'links') except: print(os.path.join('downloads',software,catalog)) raise
def to_sqlite3(): dt = DumpTruck('/tmp/open-data.sqlite', auto_commit = False) dummyrow = dict(zip(['software','catalog','identifier'], ['blah']*3)) dt.create_table(dummyrow, 'datasets', if_not_exists = True) dt.create_index(['software','catalog','identifier'], 'datasets', if_not_exists = True, unique = True) for table in ['ckan','socrata']: dt.create_table({'catalog':'blah','identifier':'blah'}, table, if_not_exists = True) dt.create_index(['catalog','identifier'], table, if_not_exists = True, unique = True) dt.create_table({'view_id':'abc','table_id':123}, 'socrata_tables') dt.create_index(['view_id'], 'socrata_tables', if_not_exists = True, unique = True) dt.create_index(['table_id'], 'socrata_tables', if_not_exists = True) for dataset in datasets(): row = { 'software': dataset['software'], 'catalog': dataset['catalog'], 'identifier': dataset[SOFTWARE_MAP['identifier'][dataset['software']]], } sql = 'SELECT * FROM datasets WHERE software = ? AND catalog = ? AND identifier = ?' if dt.execute(sql, [row['software'],row['catalog'],row['identifier']]) != []: continue dt.upsert(row, 'datasets') if dataset['software'] == 'socrata': socrata_table = { 'view_id': row['identifier'], 'table_id': dataset['tableId'], } dt.upsert(socrata_table, 'socrata_tables') dt.upsert(dataset,dataset['software']) dt.commit()
dt = DumpTruck(dbname='/tmp/finalip.db') dt.create_table({u'DA Number': u'NAE-2009-01067'}, 'finalip', if_not_exists=True) dt.create_index(['Da Number'], 'finalip', unique=True, if_not_exists=True) # Skip finished stuff pages = set([(row['Year'], row['Month'], row['Page']) for row in dt.execute('SELECT Year, Month, Page FROM finalip')]) # Populate for dirname, subdirnames, filenames in os.walk( os.path.join(os.environ['READER_ROOT'], '..', 'finalips')): if subdirnames != []: continue for filename in filenames: year, month = map(int, dirname.split('/')[-2:]) page = (year, month, filename) if page in pages: continue path = os.path.join(dirname, filename) try: data = read_finalip(path) except: print path raise for row in data: row['Year'], row['Month'], row['Page'] = page dt.upsert(data, 'finalip')
raise return map(do_row, trs[2:]) # Schema dt = DumpTruck(dbname = '/tmp/finalip.db') dt.create_table({u'DA Number': u'NAE-2009-01067'}, 'finalip', if_not_exists = True) dt.create_index(['Da Number'], 'finalip', unique = True, if_not_exists = True) # Skip finished stuff pages = set([(row['Year'], row['Month'], row['Page']) for row in dt.execute('SELECT Year, Month, Page FROM finalip')]) # Populate for dirname, subdirnames, filenames in os.walk(os.path.join(os.environ['READER_ROOT'], '..', 'finalips')): if subdirnames != []: continue for filename in filenames: year, month = map(int, dirname.split('/')[-2:]) page = (year, month, filename) if page in pages: continue path = os.path.join(dirname, filename) try: data = read_finalip(path) except: print path raise for row in data: row['Year'], row['Month'], row['Page'] = page dt.upsert(data, 'finalip')
def _diff(self, test_name, expected_rowcount): self._load_fixture_sql('fixtures/' + test_name + '.sql') dt = DumpTruck(dbname = self.dbfile) newdata = json.loads(open('fixtures/' + test_name + '.json').read()) dt.upsert(newdata) self._rowcount_should_be(expected_rowcount)
["http://www.scotlandoffice.gov.uk/scotlandoffice/16668.141.html?tID=16677&mon=jul", "Latest releases"], ["http://www.scotlandoffice.gov.uk/scotlandoffice/16668.141.html?tID=16676&mon=aug", "Latest releases"], ["http://www.scotlandoffice.gov.uk/scotlandoffice/16668.141.html?tID=16678&mon=sep", "Latest releases"], ["http://www.scotlandoffice.gov.uk/scotlandoffice/16668.141.html?tID=16679&mon=oct", "Latest releases"], ["http://www.scotlandoffice.gov.uk/scotlandoffice/16668.141.html?tID=16680&mon=nov", "Latest releases"], ["http://www.scotlandoffice.gov.uk/scotlandoffice/10804.146.html", "Archive releases"], # 2005 ["http://www.scotlandoffice.gov.uk/scotlandoffice/10805.145.html", "Archive releases"], # 2006 ["http://www.scotlandoffice.gov.uk/scotlandoffice/10806.144.html", "Archive releases"], # 2007 ["http://www.scotlandoffice.gov.uk/scotlandoffice/10807.143.html", "Archive releases"], # 2008 ["http://www.scotlandoffice.gov.uk/scotlandoffice/13342.html", "Archive releases"], # 2009 ["http://www.scotlandoffice.gov.uk/scotlandoffice/13661.html", "Archive releases"], # 2010 ["http://www.scotlandoffice.gov.uk/scotlandoffice/15263.html", "Archive releases"], # 2011 ] dt = DumpTruck(dbname="scotland.db") dt.create_table({"Title": "", "Publication date": "", "Old URL": "", "Summary": "", "Attachments": "", "Type": "", "Associated organisations": ""}, "publications") dt.create_index(["Title", "Old URL"], "publications", unique=True) for url, page_type in URLS: for publication in scrape_list_page(url): publication['Type'] = page_type dt.upsert(publication, "publications") dumptruck_to_csv(dt, "publications", "/home/http/scotland/publications.csv")
def main(): dt = DumpTruck(dbname='metrics.db') dt.create_table({'portal': 'abc', 'date': datetime.date.today()}, 'series') dt.create_index(['portal', 'date'], 'series') dt.upsert(list(table()), 'series')
"Associated Document Series": ""}, "statistics") dt.create_index(["Title", "Old URL"], "statistics", unique=True) for link in doc.xpath("//div[@class='wrapper']/ul/li/a"): series_title, series_url = link.text, urlparse.urljoin(URL, link.attrib["href"]) print series_title series_req = requests.get(series_url) series_doc = lxml.html.fromstring(series_req.text) for table_line in series_doc.xpath("//tr[not(@bgcolor) or @bgcolor!='#004093']"): file_pub_date = table_line.xpath("./td[3]")[0].text for file_node in table_line.xpath("./td[2]//a"): file_title = etree.tostring(file_node, method="text", encoding="utf8") file_link = file_node.attrib["href"] if not file_link.startswith("http"): file_link = urlparse.urljoin(URL, file_link) file_data = {"Old URL": series_url, "Title": file_title, "Body": file_title, "Publication date": datetool.parsedate(file_pub_date), "Attachment": file_link, "Attachment title": file_title, "Associated organisations": "Scotland Office", "Associated Document Series": series_title} dt.upsert(file_data, "statistics") dumptruck_to_csv(dt, "statistics", "/home/http/scotland/stats.csv")
def main(): dt = DumpTruck(dbname = 'metrics.db') dt.create_table({'portal': 'abc', 'date': datetime.date.today()}, 'series') dt.create_index(['portal', 'date'], 'series') dt.upsert(list(table()), 'series')
def scrape_main_article(url): req = requests.get(url) doc = lxml.html.fromstring(req.text) div = doc.xpath("//*[@class='wrapper']")[0] div.remove(div.find("h1")) for para in div.findall("p"): if para.find("strong") is not None: div.remove(para) return htmlize(etree.tostring(div)) dt = DumpTruck(dbname="scotland.db") dt.create_table({"Title": "", "Publication date": "", "Old URL": "", "Summary": "", "Body": "", "Associated organisations": ""}, "news") dt.create_index(["Title", "Old URL"], "news", unique=True) for url in URLS: for news_item in scrape_list_page(url): attachments = json.loads(news_item.pop("Attachments")) link = attachments[0]["link"] news_item["Old URL"] = link news_item["Body"] = scrape_main_article(link) dt.upsert(news_item, "news") dumptruck_to_csv(dt, "news", "/home/http/scotland/news.csv")
while True : idt="%s" % random.randrange(1,mid) if idt in done : sc=sc+1 else : d=dt.execute("select count(*) as c from events where id='%s'" % idt) if d[0]["c"]==1 : # logger.debug("%s already scraped" % idt) sc=sc+1 done[idt]=1 else : retry=0 success=False while success==False and retry<3 : try : dt.upsert(eventdata(idt,proxy=proxy),"events") success=True except OperationalError : retry += 1 time.sleep(1) logger.error("Retrying #%s" % retry) except KeyboardInterrupt : raise except Exception, e: logger.exception("ERROR %s " % (idt,)) err+=1 sc=sc+1 if (sc % 100)==0 : nn=scrape_now() try : stats={
'http://www.nwo.usace.army.mil', 'http://www.nws.usace.army.mil', 'http://www.nww.usace.army.mil', 'http://www.pof.usace.army.mil', 'http://www.poj.usace.army.mil', 'http://www.saw.usace.army.mil', 'http://www.spa.usace.army.mil', 'http://www.spk.usace.army.mil', 'http://www.spl.usace.army.mil', 'http://www.swf.usace.army.mil', 'http://www.swg.usace.army.mil', 'http://www.tam.usace.army.mil', } if __name__ == '__main__': dt = DumpTruck(dbname = 'usace.db') dt.create_table({'permit_application_number': 'abcd'}, 'notice') dt.create_index(['permit_application_number'], 'notice') for division in parse.locations(get('http://www.usace.army.mil/Locations.aspx')): for district in division['districts']: domain = re.sub(r'.usace.army.mil.*$', '.usace.army.mil', district['href']) path = '/Missions/Regulatory/PublicNotices.aspx' if domain in SKIPPED_DISTRICTS: continue pn_list = None while pn_list == None or pn_list['last_page'] > pn_list['current_page']: pn_list = parse.public_notice_list(get(domain + path)) dt.upsert(list(pn_list['notices']), 'notice')
from collections import defaultdict from slscraper import eventdata d=True c=1 while d: r=dt.execute("select id from events where error like '%JSON%' order by random() limit 1") if len(r)>0 : rr=r[0]["id"] try : dd=eventdata("%s" % rr) if 'error' not in dd : dd["error"]="--" dt.upsert(dd,"events") print "%(id)s %(title)s: %(error)s" % defaultdict(lambda : "-", dd) c+=1 if (c % 10)==0 : r=dt.execute("select count(*) as c from events where error like '%JSON%'") if len(r)>0 : print "still %s to go" % r[0]["c"] except Exception, e: print "ERROR %s %s" % (rr,e) else : d=False # <codecell>
#!/usr/bin/env python from dumptruck import DumpTruck dt = DumpTruck(dbname='scotland.db') dt.create_table({'title': '', 'date': ''}, 'publications') dt.create_index(['title'], 'publications', unique=True) dt.upsert({'title': 'one', 'date': 'today'}, 'publications') dt.upsert({'title': 'one', 'date': 'yesterday'}, 'publications') data = dt.execute('SELECT * FROM `publications`') print data